CN-121981324-A - Multi-source data fusion monitoring method and system for cement clinker yield

CN121981324ACN 121981324 ACN121981324 ACN 121981324ACN-121981324-A

Abstract

The invention provides a multi-source data fusion monitoring method and system for cement clinker yield, wherein the method comprises the steps of determining at least two real-time clinker yields based on at least two different clinker real-time yield accounting methods and obtained monitoring data of a target production line, determining yield differences of the at least two real-time clinker yields, generating first early warning information when the yield differences are larger than first preset differences, determining monthly clinker yields based on a material balance principle by utilizing monthly clinker sales, monthly cement sales, clinker storage period end stock, clinker storage period initial stock and outsourcing quantity, determining at least two monthly accumulated yields based on at least two real-time yield data, generating second early warning information when the difference between the monthly accumulated yields and the monthly clinker yields is larger than second preset differences, and calibrating key parameters of the clinker real-time yield accounting methods. The invention gives consideration to the real-time performance and accuracy of clinker yield.

Inventors

HU HONGBO
DENG XIKUN
MENG LINGDONG
WANG XINGPENG
YE MENG

Assignees

湖北省计量测试技术研究院
华新建材集团股份有限公司
碳排放权登记结算(武汉)有限责任公司

Dates

Publication Date: 20260505
Application Date: 20260116

Claims (10)

1. The multi-source data fusion monitoring method for cement clinker yield is characterized by comprising the following steps: Determining at least two real-time clinker yields based on at least two different clinker real-time yield accounting methods and the acquired monitoring data of the target production line; Determining a yield difference value of the yields of the at least two real-time clinker, and generating first early warning information when the yield difference value is larger than a first preset difference value; Determining the yield of the monthly clinker by using the sales amount of the monthly clinker, the sales amount of the monthly cement, the end stock amount of the clinker storage period, the initial stock amount of the clinker storage period and the outsourcing amount based on a material balance principle; And determining at least two month accumulated yields based on the at least two real-time yield data, generating second early warning information when the difference between the month accumulated yields and the month clinker yields is larger than a second preset difference, and calibrating key parameters of the clinker real-time yield accounting method.
2. The method of claim 1, wherein the real-time clinker yield accounting method comprises a raw clinker reduction factor method, a bucket elevator current method, and a bucket elevator power method, and wherein the real-time clinker yield comprises a first real-time clinker yield, a second real-time clinker yield, and a third real-time clinker yield corresponding to the raw clinker reduction factor method, the bucket elevator current method, and the bucket elevator power method, respectively.
3. The method of claim 2, wherein determining a yield difference between the at least two real-time clinker yields, and generating first warning information when the yield difference is greater than a first predetermined difference, comprises: Determining a first yield difference of the first real-time clinker yield and the second real-time clinker yield; Determining a second yield difference of the first real-time clinker yield and the third real-time clinker yield; and when the first yield difference value and/or the second yield difference value is larger than the first preset difference value, generating the first early warning information.
4. The method for multi-source data fusion monitoring of cement clinker yield according to claim 1, wherein the monthly clinker yield is: In the formula, Yield for month clinker; is clinker consumption; sales for monthly clinker; end of storage for clinker; Initial inventory for clinker storage period; Is the outsourcing quantity; sales for monthly cement; a final inventory for cement storage period; Initial stock quantity for cement storage period; is the cement clinker coefficient.
5. A method for multi-source data fusion monitoring of cement clinker yield according to claim 3, wherein the key parameters of the raw clinker coefficient method are raw clinker coefficient of folding, and the method further comprises: Acquiring a real-time value of an influence factor influencing the raw clinker folding coefficient, and inputting the real-time value into a pre-constructed dynamic relation model of the value of the influence factor and the raw clinker folding coefficient to obtain the raw clinker folding coefficient; wherein the dynamic relation model is obtained by training a clinker folding coefficient based on a historical value of an influence factor.
6. A method for multi-source data fusion monitoring of cement clinker yield according to claim 3, wherein the key parameter of the bucket elevator current method is a unit current clinker quantity, and the key parameter of the bucket elevator power method is a unit power clinker quantity.
7. The method for multi-source data fusion monitoring of cement clinker production according to claim 1, further comprising: accumulating the real-time clinker yield to obtain a first clinker annual yield, and accumulating the monthly clinker yield to obtain a second clinker annual yield; And determining an annual yield difference between the first clinker annual yield and the second clinker annual yield, and generating warning information when the annual yield difference is greater than a third preset difference.
8. The method for multi-source data fusion monitoring of cement clinker production according to claim 1, further comprising: When the yield difference value is smaller than or equal to a first preset difference value, inputting the monitoring data into a health state monitoring model to obtain the health state of the target production line; The health state monitoring model is obtained based on historical monitoring time sequence data training.
9. The method for multi-source data fusion monitoring of cement clinker production according to claim 1, further comprising: when the yield difference value is larger than a first preset difference value, a label to be checked is given to the monitoring data; When the monitoring data is data to be checked, performing pattern matching according to a historical abnormal pattern library to generate a marking suggestion containing one or more possible reasons and corresponding probabilities; responding to a mark application submitted by a user according to the mark proposal, and checking whether the production state declared by the mark application is matched with the equipment running state reflected by the monitoring data; If so, accepting the mark and updating the label of the monitoring data to be verified; if the two reasons are not matched, rejecting the mark and recording the mark application and the reasons which are not consistent as high risk checking clues.
10. A multi-source data fusion monitoring system for cement clinker production comprising: A real-time clinker yield determination unit for determining at least two real-time clinker yields based on at least two different clinker real-time yield accounting methods and the acquired monitoring data of the target production line; the real-time yield evaluation unit is used for determining a yield difference value of the at least two real-time clinker yields, and generating first early warning information when the yield difference value is larger than a first preset difference value; the monthly clinker yield determining unit is used for determining the monthly clinker yield by utilizing the monthly clinker sales amount, the monthly cement sales amount, the clinker storage period end stock amount, the clinker storage period initial stock amount and the outsourcing amount based on the material balance principle; And the key parameter calibration unit is used for determining at least two month accumulated yields based on the at least two real-time yield data, generating second early warning information when the difference between the month accumulated yields and the month clinker yields is larger than a second preset difference, and calibrating key parameters of the clinker real-time yield accounting method.

Description

Multi-source data fusion monitoring method and system for cement clinker yield Technical Field The invention relates to the technical field of clinker accounting, in particular to a multi-source data fusion monitoring method and system for cement clinker yield. Background The cement industry is an important basic industry of national economy and is also an important field of energy consumption and carbon emission. In order to achieve the aim of double carbon and promote industry fair competition and high-quality development, accurate and reliable yield monitoring based on the recorded capacity is implemented on a cement clinker production line, and the method has become a core requirement of industry supervision. Currently, clinker yield monitoring relies mainly on two technological paths: First, a real-time monitoring method based on a production field sensor. Such methods can provide high frequency (e.g., daily) throughput data, meeting the need for real-time monitoring. However, such methods have inherent drawbacks in that the sensor measurements are subject to errors and drift over time, and that the critical parameters on which the calculation model depends (e.g. raw clinker folding coefficient) are subject to misalignment due to raw material, process variations. And secondly, a monthly inventory accounting method based on a material balance principle. The method calculates the total monthly yield by counting the sales, inventory change, outsourcing and other trade and inventory data of the monthly clinker/cement and utilizing the physical conservation law. The result is based on physical measurement and financial data, has high accuracy and strong authority, but has the obvious defects of serious hysteresis and cannot be used for real-time monitoring, early warning and regulation of the production process. In other words, in the prior art, two types of methods, namely, real-time monitoring and monthly accounting, are independently operated, so as to form a fast, inaccurate, quasi-fast and fast data island. Therefore, it is needed to provide a method and a system for monitoring the multi-source data fusion of the cement clinker yield, which can organically fuse the two to construct a clinker yield monitoring system which has real-time response capability and can continuously calibrate itself by an authoritative standard so as to output long-term trusted data, thereby supporting accurate productivity supervision, carbon quota accounting and industry management. Disclosure of Invention In view of the foregoing, it is necessary to provide a method and a system for monitoring the cement clinker yield by multi-source data fusion, which are used for solving the technical problems in the prior art that the accounting efficiency and the accounting accuracy cannot be considered. In order to solve the technical problems, in a first aspect, the present invention provides a multi-source data fusion monitoring method for cement clinker yield, including: Determining at least two real-time clinker yields based on at least two different clinker real-time yield accounting methods and the acquired monitoring data of the target production line; Determining a yield difference value of the yields of the at least two real-time clinker, and generating first early warning information when the yield difference value is larger than a first preset difference value; Determining the yield of the monthly clinker by using the sales amount of the monthly clinker, the sales amount of the monthly cement, the end stock amount of the clinker storage period, the initial stock amount of the clinker storage period and the outsourcing amount based on a material balance principle; And determining at least two month accumulated yields based on the at least two real-time yield data, generating second early warning information when the difference between the month accumulated yields and the month clinker yields is larger than a second preset difference, and calibrating key parameters of the clinker real-time yield accounting method. In one possible implementation, the clinker real-time yield accounting method includes a raw clinker folding coefficient method, a bucket elevator current method, and a bucket elevator power method, and then the real-time clinker yield includes a first real-time clinker yield, a second real-time clinker yield, and a third real-time clinker yield corresponding to the raw clinker folding coefficient method, the bucket elevator current method, and the bucket elevator power method, respectively. In a possible implementation manner, the determining the yield difference of the at least two real-time clinker yields, when the yield difference is greater than a first preset difference, generates first early warning information, including: Determining a first yield difference of the first real-time clinker yield and the second real-time clinker yield; Determining a second yield difference of the first real-time clinker yield a